Abstract
Recently, matrix factorization-based data representation methods exhibit excellent performance in many real applications. However, traditional deep semi-nonnegative matrix factorization (DSNMF) models the relationship between samples by predefining a fixed graph, which is not optimal and thus cannot exploit the intrinsic local structure among data effectively. In this work, an adaptive graph regularized deep semi-nonnegative matrix factorization (AGRDSNMF) algorithm is proposed for data representation. This proposed AGRDSNMF method can construct an adaptive optimal graph in each layer, whose weights are automatically determined by the probabilities between neighborhood samples. Then the adaptive graph regularizer of each layer is adopted to constrain the corresponding coefficient matrix during decomposition. Therefore, AGRDSNMF can capture the geometric structure of the representation in each layer. Experiments are conducted on COIL20, PIE, and TDT2 datasets, and our AGRDNSMF algorithm can achieve encouraging clustering performance.
Similar content being viewed by others
References
Ma J, Zhang Y, Zhang L, Du B, Tao D (2019) Pseudo supervised matrix factorization in discriminative subspace. In: International Joint Conference on Artificial Intelligence, (IJCAI), pp 4554–4560
Nie F, Pei S, Wang R, Li X (2020) Fast clustering with co-clustering via discrete non-negative matrix factorization for image identification. In: IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), pp 2073–2077
Wang D, Gao X, Wang X, He L (2019) Label consistent matrix factorization hashing for large-scale cross-modal similarity search. IEEE Trans Pattern Anal Mach Intell 41(10):2466–2479
Zhang D, Wu X-J (2020) Scalable discrete matrix factorization and semantic autoencoder for cross-media retrieval. IEEE Trans Cybern. https://doi.org/10.1109/TCYB.2020.3032017
Wang D, Wang Q, Gao X (2018) Robust and flexible discrete hashing for cross-modal similarity search. IEEE Trans Circuits Syst Video Technol 28(10):2703–2715
Du B, Wang S, Wang N et al (2016) Hyperspectral signal unmixing based on constrained non-negative matrix factorization approach. Neurocomputing 204:153–161
Wang N, Du B, Zhang L (2013) An endmember dissimilarity constrained non-negative matrix factorization method for hyperspectral unmixing. IEEE J Sel Top Appl Earth Observ Remote Sens 6(2):554–569
Cai D, He X, Han J (2011) Locally consistent concept factorization for document clustering. IEEE Trans Knowl Data Eng 23(6):902–913
Salehani YE, Arabnejad E, Rahiche A et al (2020) MSdB-NMF: MultiSpectral document image binarization framework via non-negative matrix factorization approach. IEEE Trans Image Process 29:9099–9112
Jiao C, Gao Y, Yu N et al (2020) Hyper-graph regularized constrained NMF for selecting differentially expressed genes and tumor classification. IEEE J Biomed Health Inform 24(10):3002–3011
Wang C, Yu N, Wu M et al (2021) Dual hyper-graph regularized supervised NMF for selecting differentially expressed genes and tumor classification. IEEE/ACM Trans Comput Biol Bioinf 18(6):2375–2383
Jia Y, Liu H, Hou J, Kwong S (2021) Semisupervised adaptive symmetric non-negative matrix factorization. IEEE Trans Cybern 51(5):2550–2562
Meng Y, Shang R, Shang F et al (2020) Semi-supervised graph regularized deep NMF with Bi-orthogonal constraints for data representation. IEEE Trans Neural Networks Learn Syst 31(9):3245–3258
Lee D, Seung H et al (2001) Algorithms for non-negative matrix factorization. Adv Neural Inf Process Syst 13:556–562
Ding C, Li T, Jordan M (2010) Convex and semi-nonnegative matrix factorizations. IEEE Trans Softw Eng 32(1):45–55
Xu W, Gong Y (2004) Document clustering by concept factorization. In: Proceedings of the 27th annual international conference on Research and development in information retrieval, vol 202, pp 202–209
Cai D, He X, Han J et al (2011) Graph regularized nonnegative matrix factorization for data representation. IEEE Trans Pattern Anal Mach Intell 33(8):1548–1560
Shang F, Jiao LC, Wang F (2012) Graph dual regularization non-negative matrix factorization for co-clustering. Pattern Recogn 45:2237–2250
Shu Z, Wu X, You C et al (2020) Rank-constrained nonnegative matrix factorization for data representation. Inf Sci 528:133–146
Ghassemian R et al (2015) Spectral unmixing of hyperspectral imagery using multilayer NMF. IEEE Geosci Remote Sens Lett 12(1):38–42
Shu Z, Zhou J, Tong L, Bai X, Zhao Cet al (2015) Multilayer manifold and sparsity constrained nonnegative matrix factorization for Hyperspectral Unmixing. In: IEEE International Conference on Image Processing (ICIP), pp 1–8
Tong L, Zhou J, Qian B et al (2019) Adaptive graph regularized multilayer nonnegative matrix factorization for Hyperspectral Unmixing. IEEE J Sel Top Appl Earth Observ Remote Sens 13:434–447
Fang H, Li A, Xu H, Wang T (2018) Sparsity-constrained deep nonnegative matrix factorization for Hyperspectral Unmixing. IEEE Geosci Remote Sens Lett 15(7):1105–1109
Feng X, Li H, Li J et al (2018) Hyperspectral Unmixing using sparsity-constrained deep nonnegative matrix factorization with Total Variation. IEEE Trans Geosci Remote Sens 56(10):6245–6257
Zhao Y, Wang H, Pei J (2019) Deep non-negative matrix factorization architecture based on underlying basis images learning. In: IEEE Transactions on Pattern Analysis and Machine Intelligence, Early Access Article
Trigeorgis G, Bousmalis K, Zafeiriou S et al (2015) A deep matrix factorization method for learning attribute representations. IEEE Trans Pattern Anal Mach Intell 39(3):417–429
Schmidhuber J (2015) Deep learning in neural networks: an overview. Neural Networks 61:85–117
Krizhevsky A, Sutskever I, Hinton G (2012) ImageNet classification with deep convolutional neural networks. In: Advances in Neural Information Processing Systems
He W, Zhang H, Zhang L (2017) Total variation regularized reweighted sparse nonnegative matrix factorization for hyperspectral unmixing. IEEE Trans Geosci Remote Sens 55(7):3909–3921
Leng C, Cai G, Yu D, Wang Z (2017) Adaptive total-variation for non-negative matrix factorization on manifold. Pattern Recognit Lett 98:68–74
Shu Z, Wu X, Fan H et al (2017) Parameter-less auto-weighted multiple graph regularized nonnegative matrix factorization for data representation. Knowl-Based Syst 131:105–112
Nie F, Wang X, Huang H (2014) Clustering and projected clustering with adaptive neighbors. In: The 20th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), pp 997–986
Acknowledgements
This work was supported by the National Natural Science Foundation of China [Grant No. 61603159, 62162033, U21B2027, 61902160], Yunnan Provincial Major Science and Technology Special Plan Projects [Grant No. 202002AD080001, 202103AA080015], Yunnan Foundation Research Projects [Grant No. 202101AT070438, 202101BE070001-056], Excellent Key Teachers of QingLan Project in Jiangsu Province.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Shu, Z., Sun, Y., Tang, J. et al. Adaptive Graph Regularized Deep Semi-nonnegative Matrix Factorization for Data Representation. Neural Process Lett 54, 5721–5739 (2022). https://doi.org/10.1007/s11063-022-10882-x
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11063-022-10882-x